Three factors in the experiments need to be considered with care in order to provide consistent initial conditions for each experiments. First, the slope and the level of the channel needs to be adjusted carefully, otherwise the flow condition would be far from the uniform flow. Secondly, the temperature of the liquid have to be controlled around 17◦C to sustain refractive index matched condition during the experiment, temperature control device and procedure were needed. Thirdly, the saturation of solid grains should be the same for each runs of experiment, the traditional way to fulfill the demand is by immersing solid into liquid for a long period of time, let solid be fully saturated, but in our case the method cannot be applied because Para-cymene would gradually etch the PMMA, causing deflections on the PMMA surface, generating noises to image processing, thus we used vacuum bottle method to accelerate the process and keep saturation of solid grains the same for each runs.
The level of the flume were exam carefully before conducting experiment, bubble level sensors helped to to this job, and level could be adjusted by two screws supports fixed on the bottom of the flume, as shown in Figure 11, this procedure had to be done before each runs to guarantee the uniform flow condition.
To maintain the temperature would be not easy, the temperature had to be low enough to keep refractive index, however if the temperature was lower than dew point of the surrounding air, dew would be condensed on the surface of channel and grains exposed in the air, the quality of images filmed during the experiment would be poor. The problem could be solved partially by following steps: air conditioner would be operated before the experiment to keep room temperature as low as possible, reducing the chance of dew forming on channel. The 80 liters of liquid would be restored in the refrigerator, and poured into the reservoir just before we start the experiment. A water cooling system consisted of portable cooler and cooler fins were in operation during the experiment to maintain the low temperature of liquid. Some preparation works needs to be done before experiments.
1. All liquid would be put into the refrigerator for a day.
2. Pick the bad grain out, the PMMA grain would be damaged after series of exper-iments for repeat shearing and colliding, some serious damage, like cracked from inside, could hurt other grains and broken into pieces, and strong glare would be happened if the laser hit the bad grain.
3. Use the constant number of grains to keep the number of grains in the system the same over different runs of experiment.
For the channel boundaries, we chose three boundary types: smooth wall smooth floor (SWSF), smooth wall rough floor (SWRF) and rough wall rough floor (RWRF), and con-ducted a series of experiments by variation of total discharge for each type of boundary, the configurations are shown in Figure 1.6. For SWSF, the boundary is the smooth sur-face of channel walls and channel floor. For SWSF, we put cylinders of 7mm in diameters on the floor with center to center distance equaling to 6D (42mm), the arrangement pro-vided a condition that the organized layering of the solid grains would be interrupted every 42mm, and on the other hand the walls remained smooth. For RWRF, both walls and floor were installled with PMMA cylinders, except 1 run (check which one), the dis-tance between cylinder center is also set to 42mm, the arrangement added more roughness to the boundary and we hope to see how the flow structure responds to the boundary variations.
The final work is cleaning, a intensive wash work needed to be done just right after the experiment because para cymene is corrosive to the PMMA, if we leaved the liquid on the PMMA surface and let it dried out, the surface would become rough, this seriously influence the optical property of grains, so the grains would be washed for three times using Ultrasonic Cleaner and soap solution, the channel needs to be cleaned by dry and soft fabrics, for the SWRF and RWRF cases, the cylinders also needed to be cleaned, and more over, needed to be polished, the whole cleanup process would takes about 2 3 hours.
Figure 1.6 Three boundary conditions of the channel adopted in the experiments: a smooth wall smooth floor (SWSF for abbreviation); a smooth wall rough floor (SWRF) and c rough wall rough floor (RWRF).
After introducing the preparation and the finishing works, the experimental procedures would be explained below:
1. Set and adjust the channel transverse level. Meanwhile put all grains into the vacuum bottles, fill the bottle with para cymene, then vacuum the bottle for 3 minutes, then turn of the valve on the bottle to keep negative pressure and put it into the refrigerator to cool them down.
2. Set up the laser sheet and high speed camera for the transverse scan.
3. Take the liquid out of the refrigerator and pour the liquid into the reservoir gently, be careful, avoid eye contacted to the liquid, it hurts. some water might be on the bottom the container so we leave the last cup of para cymene inside the container.
4. Take the grains inside the vacuumed bottle out of the refrigerator, pour those grains into the channel directly.
5. Turn on the pump tho draw liquid into the channel just after the previous step, the reason is to prevent dew forming on the surface of grains.
6. Adjust the liquid discharge to the desired quantity, then make adjustment to the channel slope until the flow depth is constant over the longitudinal direction.
7. Wait for 5 minutes to make sure there is no variation on the level of free surface, then we performed the transverse scan.
8. Measure the flow depth.
9. Adjust the setting of laser sheet and camera to perform hr longitudinal scan.
10. Measure the discharge at outlet, to be careful not to touch the camera because the camera calibration have not done yet.
11. Take the grains out of the channel and circulation system, block the sill by a plate to raise the liquid level, put the calibration target into the channel and immersed fully in the liquid, then record the calibration image for longitudinal scan.
12. Turn off the pump, wait until the liquid is in still, then take some photos to capture the channel inclination.
13. Draw the liquid out of the reservoir and put them back to the refrigerator.
14. wash and clean everything.